Influence of concurrent use of conventional and E-cigarettes on rheumatoid arthritis risk among Korean adults: Findings from KNHANES 2015–2021

Abstract

Objective

Rheumatoid arthritis (RA) is a chronic autoimmune disease influenced by genetic and environmental factors, including cigarette smoking. However, the impact of e-cigarette use on RA has not been adequately investigated. This study investigated the association between conventional, and e-cigarette use and RA risk in adults aged ≥19 years.

Methods

We performed a cross-sectional analysis of data from the Korea National Health and Nutrition Examination Survey (KNHANES) between 2015 and 2021. A total of 21,931 adults aged 19 years and older were categorized into non-smokers, conventional cigarette users, and dual users. RA was defined based on self-reported physician diagnosis. Inverse probability treatment weighting (IPTW)-adjusted logistic regression was employed to estimate adjusted odds ratios (aORs) and 95 % confidence intervals (CIs) for RA.

Results

IPTW-adjusted logistic regression demonstrated that dual users had the highest risks of RA (aOR = 1.51, 95 % CI: 1.22, 1.87), while conventional cigarette users also exhibited elevated risks.

Conclusions

Dual use of conventional and e-cigarettes is associated with elevated RA risk, even among individuals with lower cumulative smoking exposure. These findings highlight the need for targeted public health strategies addressing dual users and underscore the importance of including e-cigarette use in RA risk assessments.

Highlights

• •Using both cigarettes and electronic cigarettes is linked to rheumatoid arthritis.
• •The link persists despite lower cumulative smoking exposure in dual users.
• •National survey of Korean adults from 2015 to 2021 using weighted analyses.
• •Findings support adding electronic cigarette use to arthritis risk assessments.

1. Introduction

Rheumatoid arthritis (RA) is a chronic autoimmune disorder characterized by persistent synovial inflammation, leading to joint destruction, disability, and diminished quality of life (Aletaha and Smolen, 2018). RA affects an estimated 0.5–1.0 % of the global population (Hunter et al., 2017; Kim and Sung, 2021), imposing substantial healthcare and societal burdens. In the United States alone, annual RA-related healthcare expenditures exceed $8.4 billion (Birnbaum et al., 2010). While the exact cause of RA is still unclear, genetic predisposition and environmental factors, such as smoking, have been recognized as critical contributors to its development and progression (Sugiyama et al., 2010). Cigarette smoking is considered the most significant modifiable risk factor associated with RA, influencing both disease onset and disease severity (Alfredsson et al., 2023).

Recently, the increasing prevalence of e-cigarette use as a substitute for conventional smoking has sparked significant interest in the public health community. E-cigarettes are often promoted as a safer alternative for tobacco products (Collins et al., 2019); however, the potential health risks and benefits associated with e-cigarette smoking remain controversial (Marques et al., 2021). In addition, the concurrent use of conventional and e-cigarettes can lead to increased toxin exposure and more severe respiratory issues compared to using either product individually (Goniewicz et al., 2018). Despite accumulating evidence regarding the health effects of e-cigarettes, their specific role in autoimmune conditions, such as RA, remains understudied (Tian et al., 2022).

Although the pathogenesis of RA remains unclear, several studies have suggested various mechanisms, including the induction of citrullinated peptides and epigenetic modifications. Citrullinated peptides, often found in the lung alveolar compartments of smokers, are thought to trigger autoimmune responses in patients with RA (Klareskog et al., 2006; Makrygiannakis et al., 2008). In addition, cigarette smoking can contribute to RA development through epigenetic changes, such as DNA methylation and histone modifications (Nair and Wilson, 2023). Whether similar or additive mechanisms are elicited by e-cigarette constituents has yet to be elucidated.

Against this background, this study aimed to evaluate the association between conventional cigarette use, e-cigarette use, and the prevalence of RA among adults aged 19 years or older. We utilized data from the Korea National Health and Nutrition Examination Survey (KNHANES) and applied inverse probability treatment weighting (IPTW).

2. Method

2.1. Study design and population

This cross-sectional study analyzed data from the Korea National Health and Nutrition Examination Survey (KNHANES) 2015–2021 (KNHANES VI-3, VII-1, VII-2, VII-3, VIII-1, VIII-2, and VIII-3). The KNHANES is an ongoing, nationwide health surveillance system conducted by the Korea Disease Control and Prevention Agency (KDCA), designed to assess the health and nutritional status of the Korean population. The study population included adults aged ≥19 years. Participants with missing covariate data (n = 575), smoking behavior (n = 16,647), or RA diagnosis (n = 4514) were excluded. E-cigarette only users were excluded owing to an insufficient sample size (n = 358). A total of 21,931 participants were included in the final analysis after excluding observations with missing data (Fig. 1).

Fig. 1.

Flow Diagram of Participant Selection and Exclusion Criteria for the Final Analytical Sample of Korean Adults Aged ≥19 Years, Korea National Health and Nutrition Examination Survey 2015–2021. RA: Rheumatoid arthritis; KNHANES: Korea National Health and Nutrition Examination Survey.

2.2. Exposure and outcome variables

Primary exposure variables were classified into three groups based on conventional cigarette and e-cigarette usage: (1) non-smokers, (2) conventional cigarette users, and (3) dual users of both conventional cigarettes and e-cigarettes. Conventional cigarette users were defined as those who had smoked >100 cigarettes in their lifetime and were currently smoking daily or occasionally. E-cigarette users were defined as those who had smoked e-cigarettes in their lifetime and had been smoking e-cigarettes in the past 30 days. Non-smokers were defined as those who were not conventional or e-cigarette users, whereas dual users were those who met both the criteria of conventional and e-cigarette users. The outcome variable was a self-reported physician diagnosis of RA (DM3_dg from the KNHANES questionnaire), coded as 0 (no diagnosis) or 1 (RA diagnosis).

2.3. Covariates

The key covariates included age, sex, BMI, household income (quartiles), and educational level. Age was treated as a continuous variable, and sex was coded as male (1) or female (2). BMI was calculated as weight (kg) divided by height squared (m²) and analyzed as a continuous variable. Household income was grouped into quartiles based on monthly household income. For interpretability, approximate U.S. dollar equivalents were provided using the mean 2021 exchange rate (₩1145 per USD) as a footnote: Q1 ≤ ₩1,000,000 (≈ $874), Q2 ≤ ₩2,076,900 (≈ $1815), Q3 ≤ ₩3,375,000 (≈ $2950), Q4 > ₩3,375,000 (≈ $2950). Educational level was grouped into four categories: middle school or below, high school graduate, college graduate, or postgraduate.

Additionally, a sensitivity analysis including pack-years of smoking was performed to account for cumulative tobacco exposure. However, due to extensive missing data, the model suffered from quasi-complete separation and produced unstable estimates, and pack-years were not significantly associated with RA.

2.4. Statistical analysis

Descriptive analyses were performed using frequencies and percentages for categorical variables and means and standard deviations for continuous variables. Chi-square tests were used to evaluate the association between categorical variables, and one-way ANOVA was used to compare means across groups. Logistic regression models were used to assess the association between smoking behavior and RA diagnosis. IPTW was used to balance the baseline characteristics and estimate the adjusted associations. A subgroup analysis of covariates was conducted after IPTW. To ensure the validity of the multivariable logistic regression analysis, assumptions were evaluated as follows. Multicollinearity among independent variables was assessed using the Variance Inflation Factor (VIF), with all values below five, indicating no significant multicollinearity. The linearity of the logit for continuous variables was checked using Box-Tidwell tests and was confirmed to be satisfied. Model fit was evaluated with the Hosmer-Lemeshow goodness-of-fit test, indicating adequate fit (p > 0.05). A sensitivity analysis was conducted excluding former smokers from the reference (non-smoker) category to evaluate the robustness of associations. Additionally, potential interaction effects between key variables (e.g., age and sex, smoking status and BMI) were tested by including interaction terms in the logistic regression models. All statistical analyses were performed using SAS software (version 9.4; SAS Institute, Inc., Cary, NC, USA).

2.5. Ethics approval

This study was exempted from review by the Institutional Review Board (IRB) of Korea University (approval number: IRB-2025-0002), as it involved secondary analysis of publicly available, anonymized data. Therefore, informed consent was not required.

3. Results

3.1. Unadjusted and adjusted basic characteristics of the study population

A total of 21,931 participants aged ≥19 years were included in the final analysis. Age, sex, BMI, income, and educational status differed significantly across the groups, as determined by chi-square tests for categorical variables and one-way ANOVA for continuous variables (p < 0.001 for all comparisons) (Table 1). After applying IPTW, statistically significant differences were observed when chi-square tests for categorical variables and one-way ANOVA for continuous variables were conducted (p < 0.001 for all comparisons) (Table S1). Group imbalances in age, sex, and socioeconomic factors were reduced, as reflected in the standardized mean differences (Table S2).

Table 1.

Baseline Characteristics of Korean Adults Aged ≥19 Years According to Smoking Status (Non-smokers, Exclusive Conventional Cigarette Users, and Dual Users), Korea National Health and Nutrition Examination Survey 2015–2021.^⁎

Characteristics	Smoking type			p⁎
	Non-smoker [N (%)] (N = 18,130) (82.67)	Conventional cigarette user [N (%)] (N = 3120) (14.23)	Dual user [N (%)] (N = 681) (3.11)
Age (years) a	52.9 ± 17.1	49.3 ± 15.2	36.5 ± 10.8	<0.0001
Sex				<0.0001
Men	6456 (35.6)	2577 (82.6)	564 (82.7)
Women	11,674 (64.4)	543 (17.4)	118 (17.3)
BMI, kg/m2 b	24.0 ± 3.6	24.3 ± 3.7	24.9 ± 4.0	<0.0001
Household income				<0.0001
1st quartile	3440 (18.9)	564 (18.1)	48 (7.1)
2nd quartile	4357 (24.1)	827 (26.5)	169 (24.8)
3rd quartile	4820 (26.6)	906 (29.0)	225 (33.0)
4 th quartile	5513 (30.4)	823 (26.4)	239 (35.1)
Education level				<0.0001
≤ Middle school	5093 (28.1)	632 (20.3)	18 (2.6)
High school	4749 (26.2)	1201 (38.5)	184 (27.0)
College graduate	7229 (39.9)	1192 (38.2)	447 (65.6)
Postgraduate	1059 (5.8)	95 (3.0)	32 (4.7)
Diagnosis of RA				<0.0001
No	17,752 (97.9)	3084 (98.9)	674 (99.0)
Yes	378 (2.1)	36 (1.1)	7 (1.0)

^⁎Derived from chi-squared tests for categorical variables and one-way ANOVA for continuous variables. ^a Mean ± standard deviation, ^b BMI was calculated as weight in kilograms divided by height in square meters, mean ± standard deviation. Abbreviations: BMI, body mass index; RA, rheumatoid arthritis.

3.2. Association between smoking types and RA: Pre-IPTW and IPTW-weighted analyses

In unadjusted model, conventional cigarette users exhibited higher odds of RA, although the association was not statistically significant (aOR = 1.06, 95 % CI: 0.92, 1.22) (Table 2). In contrast, dual users showed significantly lower odds of RA (aOR = 0.65, 95 % CI: 0.54, 0.77). In adjusted model, which was adjusted for all covariates, both conventional cigarette users and dual users showed higher odds of RA, with dual users presenting the greatest risk (aOR = 1.51, 95 % CI: 1.22, 1.87). In Model 1, conventional cigarette smokers had a significantly lower likelihood of RA than non-smokers (aOR = 0.55, 95 % CI: 0.39, 0.77), while dual users also displayed a non-significant reduced association (aOR = 0.49, 95 % CI: 0.23, 1.04) (Table S3). In the comprehensive Model 2, which accounted for age, sex, BMI, income level, and educational background, the association between conventional cigarette smoking and RA remained non-significant (aOR = 1.07, 95 % CI: 0.74–1.55). However, dual users continued to demonstrate a marginally increased likelihood of RA (aOR = 2.13, 95 % CI: 0.97–4.66), suggesting that the impact of dual smoking types on RA risk warrants further investigation. To assess the robustness of our findings, we performed a sensitivity analysis excluding former smokers from the non-smoker group. The results remained consistent, with current smokers showing significantly higher odds of RA compared to non-smokers (OR = 1.27, 95 % CI: 1.08, 1.50), suggesting that the main results were not driven by the inclusion of former smokers (data not shown).

Table 2.

Association Between Smoking Status (Non-smokers, Exclusive Conventional Cigarette Users, and Dual Users) and Prevalent Rheumatoid Arthritis Assessed by Inverse Probability of Treatment Weighting-Adjusted Logistic Regression Among Korean Adults Aged ≥19 Years, Korea National Health and Nutrition Examination Survey 2015–2021.^a

Smoking behavior	Unadjusted OR (95 % CI)	IPTW-adjusted OR a (95 % CI)
Non-smoker	1.00	1.00
Conventional cigarette user	1.06 (0.92, 1.22)	1.19 (1.03, 1.37)
Dual user	0.65 (0.54, 0.77)	1.51 (1.22, 1.87)

^aAdjusted model for age, sex, BMI, income status, and educational status.

3.3. Pack-years of conventional cigarette smokers and dual smokers

Conventional cigarette smokers exhibited significantly higher pack-years (20.0 ± 15.8) than dual smokers (11.2 ± 10.6). One-way ANOVA confirmed that this difference was statistically significant (p < 0.0001) (Table S4).

4. Discussion

Our findings indicate distinct differences in the association between smoking behavior and RA prevalence depending on the type of smoking and demographic characteristics. Conventional cigarette smoking initially appeared to be inversely associated with RA in the unadjusted model before IPTW; however, this relationship did not reach statistical significance after controlling for age, sex, BMI, education level, and household income. Attenuation of the protective effect after adjustment highlights the influence of demographic factors, particularly age and sex, which are strong covariates of RA risk. This suggests that the observed crude association may reflect confounding by these variables rather than a true protective effect.

Notably, in the fully adjusted model including pack-years prior to IPTW application, we observed signs of quasi-complete separation, likely due to low event counts and multicollinearity. This statistical phenomenon highlights the limitations of conventional regression models when dealing with sparse outcome data and underscores the utility of IPTW in improving model stability and validity.

After adjusting for IPTW, conventional cigarette smokers showed higher odds of RA than non-smokers. Dual users of conventional cigarettes and e-cigarettes exhibited significantly higher odds of developing RA than non-smokers after full adjustment for confounders. Interestingly, while conventional cigarette users reported a mean smoking history of 20 pack-years, dual users reported a substantially lower mean smoking history of 11 pack-years. The likelihood of developing RA typically increases with cumulative smoking exposure (Costenbader et al., 2006; Di Giuseppe et al., 2014). However, dual users exhibited a heightened risk despite fewer pack-years, suggesting possible synergistic effects of e-cigarette use on immune dysregulation and inflammatory pathways. The constituents of e-cigarettes, such as flavoring agents, heavy metals, and volatile organic compounds, are known to induce oxidative stress and enhance pro-inflammatory cytokine production, which may contribute to autoimmune activation(Qasim et al., 2017). These findings highlight the additive or synergistic risks associated with dual nicotine exposure (Reddy et al., 2021).

The group imbalances in the study population are noteworthy. E-cigarettes, a relatively new product, are less commonly used by older adults (Kramarow and Elgaddal, 2023). This age-related disparity contributes to the imbalance between the conventional and dual smoking groups. Additionally, underreporting of smoking behaviors among women may have exacerbated group disparities, reflecting sociocultural biases in self-reported smoking data (Gunter et al., 2019).

Although substantial evidence supports an elevated risk of RA among female smokers, particularly those using conventional cigarettes, there is currently a lack of robust epidemiological data specifically addressing RA risk among women who engage in dual use of conventional cigarettes and e-cigarettes. Most prior studies on dual use have focused on other health outcomes, such as serum uric acid levels or cardiovascular risk (Chen et al., 2024), rather than autoimmune diseases like RA. This gap in the literature highlights the need for future research to clarify the unique risks associated with dual tobacco product use in women. BMI ≥25 kg/m² was consistently associated with higher RA risk, aligning with studies linking obesity to inflammatory pathways (Qin et al., 2015). Socioeconomic disparities, such as lower income (Ghawi et al., 2015) and education levels (Bengtsson et al., 2005), further highlight the influence of social determinants on RA risk.

The updated methods, including the refined IPTW approach and stratified logistic regression analyses, improved the ability of the study to address selection bias and reveal data patterns. The IPTW approach effectively reduced baseline imbalances and ensured comparisons between the exposure groups. This study captured the heterogeneity in RA risk that would otherwise have been masked in the aggregate analyses, by incorporating interaction terms and stratifying them according to sex. Given the increasing prevalence of dual smoking behaviors, these results emphasize the importance of integrating smoking cessation efforts into broader RA prevention strategies(Coleman et al., 2022). The heightened risk associated with the combined use of conventional cigarettes and e-cigarettes emphasizes the need for targeted interventions to address dual-use behaviors.

Finally, the complex interplay of socioeconomic factors with smoking behavior and RA risk highlights the importance of addressing broader social determinants of health in RA prevention strategies. By prioritizing equity-focused strategies that reduce barriers to healthcare and smoking cessation, public health efforts can mitigate the disproportionate impact of RA on vulnerable populations.

This study has several limitations. First, the cross-sectional design limits causal inference and precludes the determination of temporality between smoking behavior and RA onset. Second, RA status was based on self-reported physician diagnosis without medical record validation or serologic confirmation. This reliance on self-reported data may introduce misclassification bias, potentially leading to under- or overestimation of RA prevalence. Third, the exclusion of individuals who exclusively use e-cigarettes due to a limited sample size reduces the generalizability of the findings to this specific group. Fourth, although KNHANES collects extensive information on comorbidities and health status, these were not included as covariates in the current model to avoid potential overadjustment. Future research should consider incorporating such variables to further assess the robustness of smoking-RA associations and address residual confounding. Fifth, the potential influence of unmeasured factors, such as genetic susceptibility or occupational exposure, cannot be entirely ruled out as residual confounders. Despite its methodological limitations, this study has several key strengths. It utilizes a large, nationally representative sample of 21,931 adults from KNHANES 2015–2021, providing robust statistical power and generalizability to the South Korean population. To our knowledge, it is the first and largest population-based study to demonstrate significantly elevated RA risk among dual users of conventional and e-cigarettes, a rapidly growing yet understudied group. The use of IPTW effectively balanced numerous covariates, reducing confounding, while consistent findings across subgroups and even among those with low cumulative smoking exposure further strengthen causal inference and highlight the clinical and public health relevance of assessing e-cigarette and dual use in RA risk evaluation.

Future research should prioritize longitudinal studies to establish causal pathways linking smoking behavior to RA. Additionally, exploring the biological mechanisms underlying these associations, including the roles of citrullinated peptides, oxidative stress, and epigenetic modifications, may provide deeper insights into the pathogenesis of RA. Given the growing prevalence of e-cigarette use, research examining its impact on autoimmune conditions in diverse populations is critical for shaping regulatory policies and preventive strategies.

In conclusion, this study provides updated evidence regarding the complex relationship between smoking behavior and RA prevalence, highlighting the heightened risks associated with dual use. These findings emphasize the importance of addressing dual nicotine exposure and tailoring interventions to account for sex-specific risks and socioeconomic disparities. Further research is needed to clarify the underlying mechanisms and inform public health strategies to reduce the burden of RA.

Funding sauces

This work was supported by the IITP (Institute of Information & Communications Technology Planning & Evaluation)-ICAN (ICT Challenge and Advanced Network of HRD) (IITP-2025-RS-2022-00156439) grant funded by the Korea government (Ministry of Science and ICT) and the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (RS-2024-00441029).

CRediT authorship contribution statement

SeongYoon Lee: Writing – original draft, Visualization, Validation, Software, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Hwamin Lee: Writing – review & editing, Validation, Supervision, Software, Resources, Project administration, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

Supplementary material

Data availability

Data will be made available on request.